Abstract Prostate cancer has a heterogeneous prognosis, and clinicians and researchers still lack definitive means to distinguish potentially lethal from indolent disease. Discovery of molecular biomarkers that improve the identification of indolent versus potentially lethal prostate cancer at diagnosis would enhance the ability to determine which patients would benefit more from immediate treatment and those who are better served by active surveillance. Studies of the association of gene expression with lethal prostate cancer have predominantly focused on the tumor epithelium; however, the prostate microenvironment has drawn increasing attention as a critical driver of cancer progression. Signaling factors from the microenvironment influence the epithelium to acquire properties such as increased motility, proliferation, and invasive behavior. Despite these important findings, characterization of stromal gene expression has been addressed only on a small scale. We previously identified genes associated with prostate cancer aggressiveness using expression data from pure laser-capture microdissected stroma from 25 radical prostatectomy specimens. We additionally have shown that DNA alterations in prostate cancer are associated with unique transcriptional programs in epithelial tissue. We and others have determined that several of the inherited prostate cancer risk variants function as expression Quantitative Trait Loci in tumor and normal epithelium; however, not all risk SNPs were associated with gene expression, possibly because the impact of genetic variants is tissue specific. We now extend these analyses to focus on gene expression of prostate stroma. We hypothesize that (1) a gene expression signature in stroma (alone or in combination with an epithelial gene expression signature) is predictive of prostate cancer aggressiveness and outcome; (2) prostate cancer subtypes with distinct DNA alterations are associated with gene expression in stroma; (3) prostate cancer risk variants are associated with stromal gene expression; and (4) the composition of cell types within the stroma surrounding the tumor may be predictive of prostate cancer aggressiveness and outcome. To test these hypotheses, we will perform gene expression profiling of tumor- and normal- associated stroma for ~400 prostate cancer cases (including 115 lethal cases) with existing epithelial gene expression data from the Physicians? Health Study and the Health Professionals Follow-up Study Tumor Cohort. We will also perform multiplex immunohistochemical staining of tissue microarrays consisting of 1500 prostate cancer cases (117 lethal cases) to study stroma cell type composition. In our project, we innovatively combine genetics, tumor and microenvironment biology, and biomarker development using rigorous statistical and bioinformatics approaches. Focus on adjacent stromal tissue is potentially transformative for clinical care by significantly adding to available epithelial molecular prognostic tests. An improved understanding of the role of the microenvironment in tumor initiation and progression may also potentially lead to the development of stroma-targeted therapies.